The SIAA Standard — byStraits Institute for Applied AI
Talk to us
A hiring manager and an HR colleague going through results together on a laptop

Test your staff and job candidates on real work with AI.

Each person does tasks from their own job with a live AI on screen. You see who uses AI well and safely, and what each person should learn next.

The HR console's overview of a staff AI check: 1,240 invited, 1,086 finished, 508 at Proficient or above, 63 who put company data at risk

Use it when you hire and when you train.

Hiring tests

Send our Candidate Screening Test to people applying for a role. You get a report on each candidate, with questions to ask at interview.

The interview page of a candidate's report: questions to ask, each following a task in the test, with what to listen for

Staff checks

Check the AI skills of a team or the whole company. Test again after training to see what changed.

The training plan in a staff member's report: short courses in order, each with the good practice it builds and the areas it lifts

Create a test for the roles you hire for.

  1. You chooseThe role, the industry, up to three work skills and the level.
  2. Our engine writes the testEvery question is set in that role's real work.
  3. We check every questionBefore anyone sits it, we check its facts and wording. We also check that copying the AI cannot pass it.
  4. You send itEach person gets their own link. You can read every question before it is sent.
The test summary in the HR console: a four-question interview test for insurance professionals at Proficient level, about 28 minutes

See where each department is strong and where it is weak.

The overview shows each department's results in the six areas we mark: three for creating value with AI, and three for managing its risks. The weakest results stand out in red.

Each department's results in the HR console: Acceleration, Expansion and Delegation for creating value; Safety, Reliability and Handover for managing risk

See what each person did, with the evidence.

Each result shows the level reached in each area, quotes from the person's own work with the AI, and the marker's note. Here, an analyst caught a wrong total in the AI's draft before it went into the report.

One person's result in the HR console: results by area, and the evidence that they caught the planted error in the AI's total

Train your people, and keep the test fair.

Training

Each result names what the person should work on next and suggests a course for it. Courses and books come from the Straits Institute's AI+ library.

Fairness

See the pass rate for each group of staff or candidates. We check it against the four-fifths rule: no group's pass rate should be below four-fifths of the best group's.

Each result leads to training, and a new test shows the change.

  1. The candidate screen: a task with the AI assistant beside it
    Sit the testReal work from the person's own job, with a live AI on screen.
  2. Page 1 of a sample report: the person's six scores and what to learn next
    Read the reportScores in six areas, strengths, and what to learn next.
  3. The report's training plan: short courses in order, each with the areas it lifts
    Follow the training planShort courses, starting with the lowest scores.
  4. Sit the test againThe new report shows what changed, and the next plan starts from it.

The new report starts the next round

Start with a pilot in one team.

Talk to us